Histopathological Predictors of Microsatellite Instability in Colorectal Cancer—a Tertiary Care Center Experience

Abstract

Microsatellite instability (MSI) has a therapeutic and prognostic implication in colorectal carcinomas (CRCs). It can be detected either by immunohistochemistry (IHC) or molecular studies. In developing countries, a significant proportion of the patients experience financial constraints limiting the utilization of healthcare facilities. We aimed to identify the possible clinicopathological variables which can be used as predictors of microsatellite instability in such patients. CRC cases received for MSI detection by IHC (for 1 and 1/2 years) were included. A panel of four IHC markers (anti-MLH1, anti-PMS2, anti-MSH2, and anti-MSH6) was used. Confirmation by the molecular study was recommended in all the IHC-proven MSI cases. Various clinicopathological parameters were evaluated as predictors of MSI. Microsatellite instability was detected in 40.6% (30/74) cases with MLH1 and PMS2 dual loss in 27% cases, MSH2 and MSH6 dual loss in 6.8%, loss of all four MMR proteins in 2.7%, and isolated PMS2 loss in 4.1%. MSI-H expression was shown by 36.5% cases with only 4.1% cases showing MSI-L expression. The age cut-off value to differentiate both the study groups (MSI vs MSS) was 63 years with a sensitivity of 47.7% and specificity of 86.7%. ROC curve showed an area under the curve of 0.65 (95% CI, 0.515–0.776; p-value = 0.03). On univariate analysis, age < 63 years, colon site, and absence of nodal metastasis were significantly higher in the MSI group. However, on multivariate analysis, only the age < 63 years was found to be significantly higher in the MSI group. Confirmation was molecular study could only be obtained in 12 cases and was completely concordant with MSI detection by IHC. MSI detection can be performed either by IHC or by molecular study. In this study, no histological parameter appeared to be the independent predictor of MSI status. The age < 63 years might predict the microsatellite instability, yet larger studies are needed for its validation. Thus, we recommend that IHC testing should be performed in all CRC cases.

Introduction

Colorectal carcinoma (CRC) is one of the commonest malignancies in the western world [1] and worldwide ranks second in females and third in males [2]. Though it is not among the leading malignancies in India, the occurrence is increasing with time, recorded as an increase in the highest age-adjusted annual incidence rate from 4.1 to 15.9 in males and from 5.2 to 11.4 in females [3]. CRC develops by one of the three pathways: chromosomal instability pathway (75%), microsatellite instability pathway (15%), or epigenetic gene silencing (10%) [4]. Microsatellite instability (MSI) may develop as a result of sporadic or germline (Lynch’s syndrome) mutations in DNA mismatch repair genes and rarely by sporadic methylation of hMLH1 [5].

High-level microsatellite instability (MSI-H) affects treatment response and prognosis. It has been reported that MSI-H CRCs are less responsive to 5 fluorouracil-based chemotherapy regimens [6] and have a better prognosis than microsatellite stable (MSS) tumors [7]. Thus, it has become vital to detect microsatellite instability in CRCs for planning treatment and prognostication. MSI testing can be done either by immunohistochemistry (IHC) or by polymerase-chain-reaction (PCR)-based molecular studies. A panel of four markers including anti-MLH1, anti-PMS2, anti-MSH2, and anti-MSH6 is recommended for MSI testing by IHC [8].

In this study, we aimed to identify the possible clinicopathological predictors of microsatellite instability.

Material and Methods

This was a retrospective study conducted in the department of pathology at Dayanand Medical College and Hospital, a tertiary care center for 1 and 1/2 years from October 2019 to March 2021.

The CRC cases received in the immunopathology laboratory in our department for the evaluation of microsatellite instability (MSI) were included in this study. The information regarding the patient’s demographics, tumor laterality, and the exact site of the tumor was retrieved from the hospital-based database.

Histopathological parameters using one or two representative sections from the tumor were reviewed by at least two pathologists involved in this study. The histological parameters were evaluated as the predictive factors for mismatch repair (MMR) protein expression and included subtypes of adenocarcinoma (conventional, mucinous, and signet ring cell), grade of conventional adenocarcinoma (mild, moderate, and severe), peritumoral lymphocyte response (absent, mild (≤ 1 lymphoid aggregate/high power field), moderate (2–3 lymphoid aggregates/high power field) and severe (> 3 lymphoid aggregates/high power field)), lymphovascular invasion by tumor cells, necrosis, any lesion in the adjacent colon, and pathological TNM stage which included extent of tumor invasion and lymph nodal stage. The revised histopathological evaluation was blinded to MSI results.

Inclusion Criteria

Only the resection specimens of the colon and rectum were included in the study.

Exclusion Criteria

The patients meeting any of the following criteria were excluded from the study group: (1) excisional (polypectomy) biopsy-proven cases of adenocarcinoma, (2) cases that did not have the tumor advancing edge within the slide, (3) intra-mucosal carcinoma or carcinoma in situ, and (4) resection specimens received after neoadjuvant chemoradiotherapy (NACRT) as it might affect the antigenic expression.

Immunohistochemistry (IHC) staining was done on Ventana Benchmark GX automated platform. Mouse monoclonal antibody markers by VENTANA were used, comprising anti-MLH1 (clone M1), anti-PMS2 (clone A16-4), anti-MSH2 (clone G219-1129), and anti-MSH6 (clone SP93).

IHC interpretation was done by a group of two or more pathologists. Every case was evaluated against on-slide external control (normal colonic epithelium) and positive internal control (stromal cells and lymphocytes). Any nuclear positivity was considered as intact MMR expression [9, 10]. However, any loss of expression of MMR protein was confirmed by repeating the IHC on another block, and only the cases showing loss of expression in the presence of positive internal control were reported as microsatellite instable (MSI).

Confirmation of MSI by molecular studies was recommended in all IHC-proven MSI patients. However, only 12 patients had got it done, and the MSI on IHC was found to be completely concordant with MSI on polymerase chain reaction (PCR). A panel of two mononucleotides (BAT 25 and BAT 26) and three dinucleotide repeats (D2S123, D5S346, and D17S250) were used on matched normal and tumor DNA samples, as recommended by the National Cancer Institute (NCI) for the evaluation of MSI in CRC [11]. Loss of 2 or more MMR proteins (≥ 30%) was labeled as high-level MSI (MSI-H), and loss of only one MMR protein (1–29%) was labeled as low-level MSI (MSI-L) [12].

Statistical Analysis

SPSS statistics 21 version for Microsoft Windows (Chicago, USA) was used for the statistical analysis of data. Data were described in terms of range, mean ± standard deviation (SD), frequencies (number of cases), and relative frequencies (percentage) as appropriate. The chi-square test was used for univariate analysis of clinicopathological parameters predicting MSI status. Multivariate analysis was performed by binary logistic regression to identify the independent predictor of MSI. Receiver operator curve (ROC) analysis was used to derive the cut-off point for the age to differentiate both the study groups (MSI vs MSS).

A probability value (p-value) less than 0.05 was considered statistically significant.

Results

A total of 74 cases were included in the present study which suited the study design. The clinico-demographic profile of the study population is mentioned in Table 1. The mean age of the study population was 58.2 years with ages ranging from 15 to 82 years. Colorectal carcinoma (CRC) in our study occurred more commonly in age < 63 years, male gender, and on the right side of the colon with ascending colon being the commonest site of involvement.

Table 1.

Clinico-demographic profile of study population

S no	Feature		Number	Percentage (%)
1	Age (years)	< 63	53	71.62
		> 63	21	28.38
2	Gender	Male	53	71.6
		Female	21	28.4
3	Laterality	Right	39	52.7
		Left	35	47.2
4	Site	Caecum	11	14.8
		Ascending colon	25	33.8
		Transverse colon	03	4.1
		Descending colon	10	13.5
		Sigmoid colon	15	20.3
		Rectum	10	13.5

On IHC examination, the majority of the CRC cases were found to be MSS (n = 44; 59.4%) with 30 (40.6%) cases being MSI. The immunohistochemical profile of the study population along with the spectrum of MSI is shown in Table 2. Out of the 30 MSI cases, the majority of cases (n = 20; 27%) in our study had MLH1 and PMS2 dual loss (Fig. 1). Of the cases, 36.5% showed MSI-H expression with only 4.1% cases showing MSI-L expression.

Table 2.

Immunohistochemical (IHC) profile of study population

S no		MMR protein expression	Number	Percentage (%)
1	MSI-H	MLH1 and PMS2 dual loss	20	27.0
		MSH2 and MSH6 dual loss	05	6.8
		Loss of all four MMR proteins	02	2.7
	MSI-L	Isolated PMS2 loss	03	4.1
2	MSS	Expression of all four MMR proteins	44	59.4

Fig. 1.

MMR-IHC analysis of colorectal carcinoma is depicted in the figure. IHC (Immunohistochemical stain: × 400) shows a loss of expression of the MMR proteins MLH-1 and PMS-2 and intact expression of MMR proteins MSH-2 and MSH-6 as shown in A, B, C, D respectively. Immunostaining of lymphocytes, stromal cells, and basal crypts is used to identify internal control

The histopathological characteristics of the CRC cases are tabulated in Table 3. In this study, the commonest histological patterns encountered were conventional adenocarcinoma subtype with moderate differentiation having a mild peritumoral lymphocytic response (PTL), no lymphovascular invasion (LVI), and no necrosis. There was no lesion in the adjacent colonic segment in most of the cases, and the commonest pTNM stage was pT3N0Mx.

Table 3.

Histopathological characteristics of the CRC cases

S no	Histopathological features			Number (%)
1	Type of carcinoma	Adenocarcinoma		60 (81.1%)
		Well-differentiated		08 (10.8%)
		Moderately differentiated		51 (68.9%)
		Poorly differentiated		01 (1.4%)
		Mucinous adenocarcinoma		10 (13.5%)
		Signet ring cell carcinoma		04 (5.4%)
2	Peritumoral lymphocytic response (PTL)	Mild		48 (64.8%)
		Moderate		20 (27.1%)
		Severe		06 (8.1%)
3	Adjacent colon abnormality	No abnormality		60 (81.0%)
		Adenomatous polyp		13 (17.5%)
		Hyperplastic polyp		01 (1.35%)
4	Lymphovascular invasion	Present		05 (6.7%)
		Absent		69 (93.3%)
5	Necrosis	Present		10 (13.5%)
		Absent		64 (86.5%)
6	pTstage	T1		01 (1.4%)
		T2		10 (13.5%)
		T3		59 (79.7%)
		T4a		02 (2.7%)
		T4b		02 (2.7%)
7	pNstage	N0		48 (64.8%)
		N1	N1a	13 (17.6%)
			N1b	01 (1.35%)
			N1c	01 (1.35%)
		N2	N2a	07 (9.5%)
			N2b	04 (5.4%)

The age cut-off value to differentiate both the study groups (MSI vs MSS) was 63 years with a sensitivity of 47.7% and specificity of 86.7%. ROC curve showed an area under the curve of 0.65 (95% CI, 0.515–0.776; p-value = 0.03) (Fig. 2).

Fig. 2.

Receiver operating characteristics (ROC) curve of age for differentiating MSI from MSS group (cut-off value: 63 years; area under the curve (AUC): 0.65 with 95% CI 0.515–0.776; p-value = 0.03)

The clinicopathological predictors of MSI are mentioned in Tables 4 and 5. On univariate analysis, age > 63 years, rectal malignancies, and pathological nodal metastasis were more commonly found in the MSS group with a statistically significant p-value of < 0.05. However, on multivariate analysis, age > 63 years was the only statistically significant factor associated with MSS.

Table 4.

Clinicopathological predictors of MSI (univariate analysis)

S No	Predictors	Within MSI (%)	Within MSS (%)	p-value
1	Age (> 63 years)	13.3	38.6	0.02*
2	Gender (male)	73.3	70.5	0.78
3	Site-rectum	3.3	20.5	0.04*
4	Right side of colon	56.6	50.0	0.57
5	Mucinous + signet ring cell carcinoma	20	18.1	0.84
6	pT3 + T4	80	86.4	0.44
7	pN( +)	20	45.5	0.03*
8	Lymphovascular invasion ( +)	10	4.5	0.35
9	Necrosis ( +)	16.7	11.4	0.51
10	Peritumoral lymphocytic response (severe)	10	6.8	0.62
11	Adenomatous polyp ( +)	13.3	20.5	0.54

^*Statistically significant

Table 5.

Multivariate analysis of independent clinicopathological parameter

S No	Clinico-pathological parameter	Odd ratio	95% C.I. for odd ratio		p-value
			Lower	Upper
1	Age (> 63 years)	3.712	1.037	13.279	0.044*
2	Site-rectum	8.588	0.967	76.280	0.054
3	pN( +)	2.901	0.926	9.084	0.068

^*Statistically significant

Discussion

MSI detection in CRC is essential for treatment decisions, prognostication, and Lynch syndrome recognition. It might also predict the tumor response to some chemotherapeutic drugs like platinum-based drugs [13, 14]. The Revised Bethesda Guidelines for testing colorectal tumors for microsatellite instability (MSI) were proposed by Umar et al. in 2004 [15]. However, the rate of MSI testing is variable across the world depending upon the treating clinicians and institutions. It is nearly impossible to conduct MSI testing in all CRC cases in developing countries like India because of financial constrain. We aimed this study to find out the possible clinicopathological parameters to guide MSI testing in developing countries like ours.

Nayak SS et al. [9] used a panel of 4 antibodies, i.e., anti-MLH1, anti-PMS2, anti-MSH2, and anti-MSH6, wherein the study population comprised unselected consecutive CRC cases. They reported an MSI prevalence of 22.9%. Pandey et al. [16] and Singh et al. [17] studied MSI frequency using only two IHC markers, i.e., MSH2 and MLH1, and reported MSI in 15.7% and 63.3% cases, respectively. The other two studies conducted on Indian patients by Kanth et al. [18] and Rajkumar et al. [19] reported MSI of 48.4% and 67.74% respectively with the study population composed of patients fulfilling the Bethesda criteria.

Similar to Nayak SS et al., consecutive cases of CRC were included in this study, and the MSI-IHC panel comprised a panel of four markers (anti-MLH1, anti-PMS2, anti-MSH2, and anti-MSH6 antibodies). The frequency of MSI cases was 40.6% in this study which was in agreement with the study conducted by Kanth et al. but was higher than that found by Nayak SS et al. (22.9%). Higher frequency in this study can be attributed to any of the possible reasons such as the exclusion of CRC cases which were excisional biopsy proven/post NACRT/lacked tumor advancing edges/intra-mucosal carcinoma, geographical differences, tumor biology differences etcetera. Yet, it demands larger studies incorporating study populations from throughout India to develop generalized data for the country.

Out of the various clinicopathological parameters studied, on univariate analysis, the absence of pathological nodal metastasis was significantly higher in the MSI group which was concordant with the study conducted by Nayak SS et al. Also, colon malignancies were found to be significantly higher in the MSI group in our study. However, right-sided colon involvement was statistically insignificant in our study in contrast to Nayak SS et al. However, on multivariate analysis, age < 63 years was the only statistically significant factor associated with MSI in our study which is in contrast to the study by Nayak SS et al. who found age < 60 years, right‑sided colonic location, and the presence of severe intratumoral lymphocyte as significant independent variables for MSI. These observations might point toward the possible role of a few of the clinicopathological parameters in estimating the loss of MMR protein expression yet needs extensive research and validation by larger population-based studies. Hence, MMR protein expression detection must be carried out in all CRC cases until a generalizable guideline of clinicopathological predictors for MSI has been established [20].

MSI testing can be done by immunohistochemistry and/or molecular studies. IHC is an indirect method that detects the expression of MMR proteins coded by MMR genes, whereas molecular studies detect the phenotypic variation in MMR genes. In our study, IHC was used as the primary detection modality, and molecular studies were recommended for the patients who were found to be MMR deficient on IHC. However, out of 30 MMR-deficient patients, only 12 patients got the molecular studies done because of financial constraints, and the results of IHC were 100% concordant with that of the molecular studies wherever performed. MLH1 and PMS2 dual loss was the commonest MSI pattern found in our study, followed by MSH2 and MSH6 dual loss and an isolated PMS2 loss. These findings were similar to those established by Nayak SS et al. However, in contrast to the study conducted by Nayak SS et al., we discovered the loss of all four MMR proteins in two cases (2.7%).

Neither IHC nor molecular study is 100% accurate when used independently. However, excellent concordance had been documented in the literature with a sensitivity of 95% and specificity of 98% by a few authors [12, 21]. The possible causes of discordance may be non-functioning MMR genes, antibody panels lacking the products of abnormal MMR genes, tumor heterogeneity, or antigen degradation [22]. The sensitivity of IHC for MSI detection has been improved by using a panel of four anti-MMR antibodies (anti-MLH1, anti-PMS2, anti-MSH2, and anti-MSH6 antibodies) over a panel of two anti-MMR antibodies (MLH1 and MSH2) [23]. IHC detection of MMR expression is a quick, simple, and cost-effective approach with good sensitivity and specificity. Internal control is also present while reporting, which is an advantage. MMR-IHC, on the other hand, has several drawbacks. To begin, heterogeneous immunostaining on MMR-IHC is a prevalent issue that can be caused by a variety of technical and biological variables. Fixation in a suitable amount of 10% neutral buffered formalin, properly thin sections, and, if necessary, a repeat IHC on a different section may all help to overcome this difficulty. Another disadvantage is false-immunopositivity caused by missense MLH1 and MSH6 mutations, which result in a shortened catalytically inactive protein that retains antigenicity [24, 25]. Aberrant staining patterns, such as cytoplasmic, dot-like, or perinuclear staining, may also be seen leading to indeterminate IHC results [23]. Post-neoadjuvant therapy related to decreased or absent MMR protein expression may occur, and in such situations, MMR-IHC on pre-treatment sample is recommended [26]. According to the European Society for Medical Oncology (ESMO) guidelines on MSI testing, MSI-PCR testing should be performed when MMR-IHC results are indeterminate, for instance, difficulties in interpreting IHC results, especially in the situation of a single heterodimer component being lost (e.g., only MLH1 or only PMS2 and not both) [27].

Sporadic versus genetic causes of MSI needs to be ascertained in all MSI cases (proven by MMR IHC and/or PCR) by molecular studies that are crucial for disease risk stratification, optimal treatment, and family screening [26]. There is abundant evidence in the literature supporting a better stage-adjusted survival in MSI-H non-metastatic CRC cases. Recommendations are that stage II CRC cases with MSI-H status be treated with surgical resection only with no adjuvant chemotherapy. However, adjuvant chemotherapy needs to be administered in all stage III and IV CRC cases irrespective of their MMR status [28]. In this study, of 30 MSI patients, 3 were stage I, 17 were stage II, and 10 were stage III patients respectively. No adjuvant chemotherapy was given in stage I (n = 3) and stage II MSI-high (n = 8) patients. However, stage II-MSS (n = 8) and stage II-MSI-low (n = 1) patients were treated with adjuvant chemotherapy comprising oral capecitabine or 5-fluorouracil. All stage III patients (n = 10) were given the FOLFOX regime (folinic acid, 5-fluorouracil, and oxaliplatin) or CAPOX regime (capecitabine, and oxaliplatin).

This study was an attempt to find clinicopathological parameters that can predict MSI in colorectal carcinomas. However, there were a few limitations to this study. First, it was performed on a small study population of a single healthcare center. Second, not all MSI cases were subjected to molecular testing. Third, follow-up was not available.

Conclusion

MSI testing in colorectal carcinomas can be performed using immunohistochemistry or molecular studies. As financial constraints continue to be a major issue in developing countries such as India, less expensive diagnostic and therapeutic modalities are required for better healthcare in limiting settings. We recommend that MSI-IHC be performed on all CRC cases for optimal treatment as no histological parameter appeared to be the independent predictor of MSI status in this study. The age < 63 years might predict the microsatellite instability, yet larger studies are needed for its validation.

Author Contribution

Aminder Singh: Concept and manuscript editing.

Saveena Jindal: Data acquisition and manuscript review.

Ankita Soni (corresponding author and guarantor): Literature search and manuscript preparation.

Harpreet Kaur: Manuscript review.

Kunal Jain: Data acquisition and manuscript editing.

Vikram Narang: Data acquisition and manuscript review.

Sumit Grover: Manuscript review.

Bhavna Garg: Manuscript editing.

Ramneek kaur: Data acquisition.

Declarations

Conflict of Interest

The authors declare no competing interests.